platform/upstream/libvpx.git
11 years agortcd: make include guard unique
James Zern [Sat, 23 Feb 2013 00:03:40 +0000 (16:03 -0800)]
rtcd: make include guard unique

prior numbers were removed resulting in the same include guard for
vp[89]_rtcd

Change-Id: If2741e5b0cbdda7390c20e865e9f05992caff93e

11 years agoMerge "Fix variance (signed integer) overflow"
James Zern [Fri, 22 Feb 2013 23:33:20 +0000 (15:33 -0800)]
Merge "Fix variance (signed integer) overflow"

11 years agoMerge "vp8: make gf_group_bits 64-bit"
James Zern [Fri, 22 Feb 2013 23:31:28 +0000 (15:31 -0800)]
Merge "vp8: make gf_group_bits 64-bit"

11 years agoMerge "vp8_first_pass(): avoid floating point div by 0"
James Zern [Fri, 22 Feb 2013 23:30:14 +0000 (15:30 -0800)]
Merge "vp8_first_pass(): avoid floating point div by 0"

11 years agovp8: make gf_group_bits 64-bit
James Zern [Fri, 22 Feb 2013 19:46:34 +0000 (11:46 -0800)]
vp8: make gf_group_bits 64-bit

avoids signed integer overflow; matches kf_group_bits

Change-Id: I193145cdc4fa53e70fba0a1731a03eb1a574931d

11 years agovp8_first_pass(): avoid floating point div by 0
James Zern [Fri, 22 Feb 2013 19:42:09 +0000 (11:42 -0800)]
vp8_first_pass(): avoid floating point div by 0

Change-Id: Id1e6a12db6b0c1d3f64ead8fd8834aadc30fbed2

11 years agoFix variance (signed integer) overflow
James Zern [Fri, 22 Feb 2013 19:28:48 +0000 (11:28 -0800)]
Fix variance (signed integer) overflow

based on change made in experimental:
  9847344 Fix variance (signed integer) overflow

Change-Id: I36f4ba5700f6f4615057daf7e70868f68a86669f

11 years agoMerge "Fixed the buffer overflow issue"
James Zern [Fri, 22 Feb 2013 20:36:32 +0000 (12:36 -0800)]
Merge "Fixed the buffer overflow issue"

11 years agoMerge "sixtap_predict_test: fix sizes passed to memset"
James Zern [Fri, 22 Feb 2013 20:32:07 +0000 (12:32 -0800)]
Merge "sixtap_predict_test: fix sizes passed to memset"

11 years agoFixed the buffer overflow issue
Jingning Han [Wed, 20 Feb 2013 23:32:17 +0000 (15:32 -0800)]
Fixed the buffer overflow issue

The issue that potentially broke the encoding process was due to the fact
that the length of token link is calculated from the total number of tokens
coded, while it is possible, in high bit-rate setting, this length is
greater than the buffer length initially assigned to the cpi->tok.

This patch increases the initially allocated buffer length assigned to
cpi->tok from
(mb_rows * mb_cols * 24 * 16) to (mb_rows * mb_cols * (1 + 24 * 16)).

It resolves the buffer overflow problem.

Change-Id: I8661a8d39ea0a3c24303e3f71a170787a1d5b1df

11 years agosixtap_predict_test: fix sizes passed to memset
James Zern [Fri, 22 Feb 2013 19:29:47 +0000 (11:29 -0800)]
sixtap_predict_test: fix sizes passed to memset

src_/dst_/dst_c_ are heap allocated, use the allocation size rather than
sizeof(var)

Change-Id: I3335ad487dc9b154cdf212891d1d74c812eff060

11 years agoMerge "Remove 'local' variables in rtcd.sh"
John Koleszar [Fri, 22 Feb 2013 16:22:33 +0000 (08:22 -0800)]
Merge "Remove 'local' variables in rtcd.sh"

11 years agoAppend .exe to the executable on OS/2
KO Myung-Hun [Fri, 22 Feb 2013 03:35:18 +0000 (12:35 +0900)]
Append .exe to the executable on OS/2

Change-Id: I1c3b17450b08d2fdfacb7b94055da29d4d20f9c8

11 years agoRemove 'local' variables in rtcd.sh
John Koleszar [Thu, 21 Feb 2013 19:11:13 +0000 (11:11 -0800)]
Remove 'local' variables in rtcd.sh

These aren't required to be supported by sh.

Change-Id: I54d641a2377531e0b320676d29d4ecdeee82288a

11 years agoUse dq instead of ddq with NASM
KO Myung-Hun [Thu, 7 Feb 2013 13:03:11 +0000 (22:03 +0900)]
Use dq instead of ddq with NASM

Change-Id: Iffb7cd44b449dc10fa5c24405be909d051b7abb5

11 years agoUse smartalign for long nops with NASM
KO Myung-Hun [Sun, 3 Feb 2013 12:44:56 +0000 (21:44 +0900)]
Use smartalign for long nops with NASM

'CPU amdnop' is supported by YASM only.

Change-Id: Ia3f7c2ba6d3bdf2889b62f5c6127fd515d7c7394

11 years agoDisable USE_POSIX_MAP on OS/2
KO Myung-Hun [Sun, 3 Feb 2013 12:43:55 +0000 (21:43 +0900)]
Disable USE_POSIX_MAP on OS/2

Change-Id: Ib88ab619fa4e1593e85ca325555f2c4648ac9bc7

11 years agoMerge "Add support for x64 and win64 yasm flags."
John Koleszar [Fri, 1 Feb 2013 01:05:33 +0000 (17:05 -0800)]
Merge "Add support for x64 and win64 yasm flags."

11 years agoAdd support for x64 and win64 yasm flags.
Frank Galligan [Thu, 31 Jan 2013 23:36:55 +0000 (15:36 -0800)]
Add support for x64 and win64 yasm flags.

Some projects must define only win64 for Windows 64bit builds using
yasm.

Change-Id: I1d09590d66a7bfc8b4412e1cc8685978ac60b748

11 years agoFix for divide by zero in vp8_adjust_key_frame.
Marco Paniconi [Thu, 31 Jan 2013 18:46:39 +0000 (10:46 -0800)]
Fix for divide by zero in vp8_adjust_key_frame.

Change-Id: I3bf9bdd95abfd287fbcb644f4fb85fb9204be95a

11 years agoMerge "WIP: Multiple decoder instances support"
Jim Bankoski [Thu, 31 Jan 2013 02:59:55 +0000 (18:59 -0800)]
Merge "WIP: Multiple decoder instances support"

11 years agoWIP: Multiple decoder instances support
Scott LaVarnway [Wed, 3 Oct 2012 17:07:13 +0000 (10:07 -0700)]
WIP: Multiple decoder instances support

Started adding support for multiple internal decoder instances.  Also added
code to limit the vp8 config options available when using frame-based
multithreading.

Change-Id: I0f1ee7abcfcff59204f50162e28254b8dd6972eb

11 years agolibvpx: Fix warnings on windows.
Frank Galligan [Tue, 29 Jan 2013 00:59:10 +0000 (16:59 -0800)]
libvpx: Fix warnings on windows.

Warnings found when tyring to build libvpx in Chromium.

Change-Id: I5824d9e2c06351e0cf46e9f5fa102cc8b04cf963

11 years agoMerge "Use FRAGMENT_DATA struct in pbi"
Scott LaVarnway [Tue, 29 Jan 2013 21:42:54 +0000 (13:42 -0800)]
Merge "Use FRAGMENT_DATA struct in pbi"

11 years agoUse FRAGMENT_DATA struct in pbi
Scott LaVarnway [Tue, 29 Jan 2013 18:34:35 +0000 (10:34 -0800)]
Use FRAGMENT_DATA struct in pbi

for fragment information.

Change-Id: Idc83625591a1e4ca6f551dcfb7fc0428f6f37351

11 years agoobj_int_extract.bat is not a generated file
Johann [Tue, 29 Jan 2013 02:19:17 +0000 (18:19 -0800)]
obj_int_extract.bat is not a generated file

Trying to create Visual Studio project files would fail with:
make[1]: *** No rule to make target `obj_int_extract.bat', needed by
`.projects'.  Stop.

Change-Id: Ie55458427ddea199a3de9973eaf2a37f711f839e

11 years agoMerge "Bug fix: error-concealment enabled changed postproc output"
Jim Bankoski [Sat, 26 Jan 2013 03:23:55 +0000 (19:23 -0800)]
Merge "Bug fix:  error-concealment enabled changed postproc output"

11 years agoBug fix: error-concealment enabled changed postproc output
Scott LaVarnway [Sat, 26 Jan 2013 01:06:32 +0000 (17:06 -0800)]
Bug fix:  error-concealment enabled changed postproc output

When error concealment is enabled, it swaps the mi and prev_mi ptrs after
each frame is decoded.  The postproc uses the mi ptr for the mode info context.
Now the postproc will use the correct mode info context.

Change-Id: I537ae5450f319c624999b44525bb52bb30047b7b

11 years agoDelay SSE check until the compiler is configured.
Johann [Sat, 26 Jan 2013 00:43:09 +0000 (16:43 -0800)]
Delay SSE check until the compiler is configured.

Move SSE4 detection below compiler configuration. Otherwise ${CC} may
not be populated.

We are specifically interested in -msse4.1. Narrow detection code.

Change-Id: I0ffac49e91b3fdb2aceb6eb6b90dc1c10b159111

11 years ago64-bit Mac Chromium support for libvpx.
Mark Mentovai [Fri, 25 Jan 2013 23:02:52 +0000 (18:02 -0500)]
64-bit Mac Chromium support for libvpx.

For 64-bit Mac Chromium, use private_extern for HIDDEN_DATA, the same as
32-bit Mac Chromium.

Change-Id: Ica0fa9e48a47409facece691ae1e39327369083c

11 years agoMerge "Include apple arm conversion script"
Johann [Fri, 25 Jan 2013 20:46:25 +0000 (12:46 -0800)]
Merge "Include apple arm conversion script"

11 years agoInclude apple arm conversion script
Johann [Fri, 25 Jan 2013 20:11:19 +0000 (12:11 -0800)]
Include apple arm conversion script

Fix building iOS targets from 'make dist' sources

Change-Id: I0bb9ba2a6d94e57c6cf439a79419af254e4a359c

11 years agoMerge "Bug fix: Handle input data ptr = NULL and size = 0 correctly"
Marco Paniconi [Fri, 25 Jan 2013 20:05:13 +0000 (12:05 -0800)]
Merge "Bug fix:  Handle input data ptr = NULL and size = 0 correctly"

11 years agoMerge "make: support V=1 as short form of verbose=yes"
James Zern [Wed, 23 Jan 2013 22:47:30 +0000 (14:47 -0800)]
Merge "make: support V=1 as short form of verbose=yes"

11 years agoBug fix: Handle input data ptr = NULL and size = 0 correctly
Scott LaVarnway [Tue, 22 Jan 2013 18:52:29 +0000 (10:52 -0800)]
Bug fix:  Handle input data ptr = NULL and size = 0 correctly

Issue 517: Issues decoding VPX_CODEC_USE_INPUT_FRAGMENTS
http://code.google.com/p/webm/issues/detail?id=517

Change-Id: I030c4cf15b1e1b993433571b6ee77c959a368ff2

11 years agoMerge "libvpx: Fix some warnings."
Jim Bankoski [Tue, 22 Jan 2013 15:58:15 +0000 (07:58 -0800)]
Merge "libvpx: Fix some warnings."

11 years agoLink pthread when it is available
Johann [Fri, 18 Jan 2013 19:31:22 +0000 (11:31 -0800)]
Link pthread when it is available

Linking when we don't use it but it is available is probably harmless.

Gtest requires pthreads. Don't automatically enable unit tests if we
don't have it.

Change-Id: I5e6c3b609f840c7b6dbb36fc65809f0ef84685f8

11 years agolibvpx: Fix some warnings.
Frank Galligan [Fri, 18 Jan 2013 17:44:23 +0000 (09:44 -0800)]
libvpx: Fix some warnings.

Change-Id: If8be8b9d28a29631f29c46daea8a226ab3580610

11 years agoDon't include x86inc.asm on non-x86 targets
John Koleszar [Thu, 17 Jan 2013 00:44:33 +0000 (16:44 -0800)]
Don't include x86inc.asm on non-x86 targets

This file is currently unused, as the asm that depended on it has been
disabled for the current roll into Chromium. It's expected that it
will return in some form, so wrap it in an x86 check rather than
deleting it. This extra file isn't really an issue with the libvpx
build system, but affects the gyp builds since on ARM (android) it
tries to do the ADS->GAS conversion on all .asm files reported in
libvpx_srcs.txt.

Change-Id: I080fbc22ab59fff41264ae230a48fa753594135b

11 years agoMerge branch 'experimental' of review:webm/libvpx
John Koleszar [Tue, 15 Jan 2013 00:25:26 +0000 (16:25 -0800)]
Merge branch 'experimental' of review:webm/libvpx

Change-Id: Ib2c2236349c2ae8ee81bd01c5067dddcbac713ca

11 years agofix a number issues that cause failures
Yaowu Xu [Tue, 15 Jan 2013 02:03:34 +0000 (18:03 -0800)]
fix a number issues that cause failures

During master jenkins verification proces

Change-Id: I3722b8753eaf39f99b45979ce407a8ea0bea0b89

11 years agoUse INT64_MAX instead of LLONG_MAX
John Koleszar [Mon, 14 Jan 2013 19:49:30 +0000 (11:49 -0800)]
Use INT64_MAX instead of LLONG_MAX

These variables have the type int64_t, not long long. long long could
be a larger type than 64 bits. Emulate INT64_MAX for older versions of
MSVC, and remove the unreferenced vpx_ports/vpxtypes.h

Change-Id: Ideaca71838fcd3849d816d5ab17aa347c97d03b0

11 years agoRemove compound intra-intra experiment.
Ronald S. Bultje [Mon, 14 Jan 2013 22:37:53 +0000 (14:37 -0800)]
Remove compound intra-intra experiment.

This experiment gives little gains and adds relatively much code
complexity (and it hinders other experiments), so let's get rid of
it.

Change-Id: Id25e79a137a1b8a01138aa27a1fa0ba4a2df274a

11 years agoMerge experiment "subpelrefmv"
Yaowu Xu [Mon, 14 Jan 2013 22:39:54 +0000 (14:39 -0800)]
Merge experiment "subpelrefmv"

Change-Id: Iac7f3d108863552b850c92c727e00c95571c9e96

11 years agoMerge experiment "widerlpf"
Yaowu Xu [Mon, 14 Jan 2013 20:11:06 +0000 (12:11 -0800)]
Merge experiment "widerlpf"

Change-Id: I0c94475075e66e13cfe4c20fab7db6474441ae86

11 years agoMerge "changed UV plane loop filtering for TX_8X8" into experimental
Yaowu Xu [Mon, 14 Jan 2013 22:47:31 +0000 (14:47 -0800)]
Merge "changed UV plane loop filtering for TX_8X8" into experimental

11 years agoMerge "change to evaluate reference mvs using above only" into experimental
Yaowu Xu [Mon, 14 Jan 2013 22:40:38 +0000 (14:40 -0800)]
Merge "change to evaluate reference mvs using above only" into experimental

11 years agochanged UV plane loop filtering for TX_8X8
Yaowu Xu [Mon, 14 Jan 2013 17:28:35 +0000 (09:28 -0800)]
changed UV plane loop filtering for TX_8X8

In commit 9a1d73d, loop filtering was added for UV 4x4 boundaries
when TX_8X8 is used by a MB. This commit further refined the decision
to be based on the actual transform used for the UV planes. When
UV planes use 4x4 transform, i.e. when prediction mode used is either
I8X8_PRED or SPLITMV, UV planes are filtered on 4x4 boundaries, and no
filtering is applied on 4x4 block boundaries when UV planes use 8X8
transform.

Change-Id: Ibb404face0a1d129b4b4abaf67c55d82e8df8bec

11 years agoMerge "Fix compiler warnings" into experimental
Paul Wilkins [Mon, 14 Jan 2013 22:20:57 +0000 (14:20 -0800)]
Merge "Fix compiler warnings" into experimental

11 years agoMerge "Merge prediction filter" into experimental
Adrian Grange [Mon, 14 Jan 2013 22:18:21 +0000 (14:18 -0800)]
Merge "Merge prediction filter" into experimental

11 years agochange to evaluate reference mvs using above only
Yaowu Xu [Thu, 10 Jan 2013 23:38:29 +0000 (15:38 -0800)]
change to evaluate reference mvs using above only

Change-Id: Ibcc342efac0a9be7a21d9b2c09984d9e16bbb225

11 years agoFix compiler warnings
Yaowu Xu [Mon, 14 Jan 2013 21:56:52 +0000 (13:56 -0800)]
Fix compiler warnings

The warnings caused verify failure with gerrit for several  commits

Change-Id: I030df8638bd69b8783a3ac58e720ff9f0bfd546c

11 years agoMerge prediction filter
Adrian Grange [Mon, 14 Jan 2013 20:57:07 +0000 (12:57 -0800)]
Merge prediction filter

Removed the experimental flag from around the prediction filter.

Change-Id: Ic1dd2db8fe8ac17ed5129f83094d4c5cdd5527d2

11 years agoReset x->skip for each iteration in the RD loop.
Ronald S. Bultje [Mon, 14 Jan 2013 20:43:12 +0000 (12:43 -0800)]
Reset x->skip for each iteration in the RD loop.

This prevents ill-defined behaviour, such as setting x->skip for a mode
that is excluded because of frame-level flags (e.g. filter selection,
compound prediction selection), then not breaking out of the RD loop
because the mode is not allowed, but keeping the flag on. Whatever mode
is iterated through next in the RD loop will then carry this flag, and
all sort of bad stuff happens, such as x->skip being set on intra pred
modes.

Change-Id: I5bec46b36e38292174acb1c564b3caf00a9b4b9a

11 years agoFix unused variable warnings
John Koleszar [Mon, 14 Jan 2013 20:10:11 +0000 (12:10 -0800)]
Fix unused variable warnings

Previous commit does not build cleanly on Jenkins with the DWT/DCT
hybrid experiment enabled (--enable-dwtdcthybrid).

Change-Id: Ia67e8f59d17ef2d5200ec6b90dfe6711ed6835a5

11 years agoconfigure: add support for android x86
Andoni Morales Alastruey [Mon, 14 Jan 2013 11:25:52 +0000 (12:25 +0100)]
configure: add support for android x86

Change-Id: Iadbfce5ecddd53a5238c2036e2c07f05d5853a5a

11 years agoFurther enhancements/fixes on dct/dwt hybrid txfm
Deb Mukherjee [Wed, 9 Jan 2013 14:26:54 +0000 (06:26 -0800)]
Further enhancements/fixes on dct/dwt hybrid txfm

Fixes some scaling issues. Adds an option to only compute the
dct on the low-low subband for 32x32 and 64x64 blocks using
only a single 16x16 dct after 1 and 2 wavelet decomposition
levels respectively. Also adds an option to use a 8x8 dct
as building block.

Currenlty with the 2/6 filter and with a single 16x16 dct on
the low low band, the reuslts compared to full 32x32 dct is
as follows:
derf: -0.15%
yt: -0.29%
std-hd: -0.18%
hd: -0.6%
These are my current recommended settings, since the 2/6 filter
is very simple.

Results with 8x8 dct are about 0.3% worse.

Change-Id: I00100cdc96e32deced591985785ef0d06f325e44

11 years agoMerge "WIP: Added sse2 version of vp9_mb_lpf_horizontal_edge_w" into experimental
Jim Bankoski [Sat, 12 Jan 2013 01:15:41 +0000 (17:15 -0800)]
Merge "WIP: Added sse2 version of vp9_mb_lpf_horizontal_edge_w" into experimental

11 years agoWIP: Added sse2 version of vp9_mb_lpf_horizontal_edge_w
Scott LaVarnway [Sat, 12 Jan 2013 01:11:04 +0000 (17:11 -0800)]
WIP: Added sse2 version of vp9_mb_lpf_horizontal_edge_w

and vp9_mb_lpf_vertical_edge_w_sse2.  This was quickly done so we can
run some tests over the weekend.  Future commits will optimize/refactor these
functions further.

The decoder performance improved by ~17% for the clip used.

Change-Id: I612687cd5a7670ee840a0cbc3c68dc2b84d4af76

11 years agoMerge "Upstream changes from Chromium Android Clang build." into experimental
Jim Bankoski [Sat, 12 Jan 2013 01:06:26 +0000 (17:06 -0800)]
Merge "Upstream changes from Chromium Android Clang build." into experimental

11 years agoMerge "Add loop filtering for UV plane" into experimental
Yaowu Xu [Sat, 12 Jan 2013 00:56:39 +0000 (16:56 -0800)]
Merge "Add loop filtering for UV plane" into experimental

11 years agoAdd loop filtering for UV plane
Yaowu Xu [Fri, 11 Jan 2013 23:53:32 +0000 (15:53 -0800)]
Add loop filtering for UV plane

On block boundary within a MB when 8x8 block boundary only is filtered
for Y.

Change-Id: Ie1c804c877d199e78e2fecd8c2d3f1e114ce9ec1

11 years agoUpstream changes from Chromium Android Clang build.
Frank Galligan [Fri, 11 Jan 2013 23:34:05 +0000 (15:34 -0800)]
Upstream changes from Chromium Android Clang build.

See https://codereview.chromium.org/11875006/

Change-Id: Ied2a17df2b3222635f84aef120eaa9feb53750d2

11 years agoMerge "Initial sse2 version of the wide loopfilters" into experimental
Scott LaVarnway [Fri, 11 Jan 2013 23:34:26 +0000 (15:34 -0800)]
Merge "Initial sse2 version of the wide loopfilters" into experimental

11 years agoInitial sse2 version of the wide loopfilters
Scott LaVarnway [Fri, 11 Jan 2013 22:54:14 +0000 (14:54 -0800)]
Initial sse2 version of the wide loopfilters

Updated the rtcd_defs and used the sse2 uv version
of the loopfilter.  The performance improved by ~8%
for the test clip used.

Change-Id: I5a0bca3b6674198d40ca4a77b8cc722ddde79c36

11 years agoRemove INT64_MAX references.
Paul Wilkins [Fri, 11 Jan 2013 19:41:39 +0000 (19:41 +0000)]
Remove INT64_MAX references.

Replace INT64_MAX references with LLONG_MAX
for windows build.

Change-Id: Ib8b45c1e9c15c043b2f54c27ed83b8682b2be34f

11 years agoMerge "Reduce the usage of widerlpf" into experimental
Yaowu Xu [Fri, 11 Jan 2013 19:15:43 +0000 (11:15 -0800)]
Merge "Reduce the usage of widerlpf" into experimental

11 years agortcd for new wider loop filters
Jim Bankoski [Fri, 11 Jan 2013 17:45:45 +0000 (09:45 -0800)]
rtcd for new wider loop filters

Change-Id: I8826bcdcf72ba6d86bde31cd13902a710399805c

11 years agoReduce the usage of widerlpf
Yaowu Xu [Thu, 10 Jan 2013 18:55:07 +0000 (10:55 -0800)]
Reduce the usage of widerlpf

The commit changed to not to use wider lpf within a superblock when
32x32 transform is used for the block.

The commit also changed to use the shorter version of loop filtering:
for UV planes.

Change-Id: I344c1fb9a3be9d1200782a788bcb0b001fedcff8

11 years agoMerge "configure: support hardfloat armv7 CHOSTS"
Johann [Thu, 10 Jan 2013 18:56:28 +0000 (10:56 -0800)]
Merge "configure: support hardfloat armv7 CHOSTS"

11 years agoMerge "configure: support mingw-w64"
Johann [Thu, 10 Jan 2013 18:35:22 +0000 (10:35 -0800)]
Merge "configure: support mingw-w64"

11 years agoFix wrong pitch argument in dct32x32 unit test.
Ronald S. Bultje [Thu, 10 Jan 2013 16:36:42 +0000 (08:36 -0800)]
Fix wrong pitch argument in dct32x32 unit test.

Change-Id: Id9474a1686daebfa3d004e21823bf1888ec9e534

11 years agoMerge tx32x32 experiment.
Ronald S. Bultje [Thu, 10 Jan 2013 16:23:59 +0000 (08:23 -0800)]
Merge tx32x32 experiment.

Change-Id: I615651e4c7b09e576a341ad425cf80c393637833

11 years agoMerge "Merge superblocks64 experiment." into experimental
Ronald S. Bultje [Thu, 10 Jan 2013 16:18:33 +0000 (08:18 -0800)]
Merge "Merge superblocks64 experiment." into experimental

11 years agoconfigure: support hardfloat armv7 CHOSTS
Luca Barbato [Thu, 10 Jan 2013 04:13:57 +0000 (05:13 +0100)]
configure: support hardfloat armv7 CHOSTS

Many linux distribution for arm switched to armhf/hardfloat, support
them.

Change-Id: I8ab39d34782b4f24b7028ac76342118166fd5905

11 years agoconfigure: support mingw-w64
Luca Barbato [Thu, 10 Jan 2013 03:50:18 +0000 (04:50 +0100)]
configure: support mingw-w64

Autodetect common cross compile and native mingw-w64 CHOSTs

Change-Id: Ib307b2f047c024c4dbc41880bd8d84b0fa7ecb93

11 years agoChange some settings in multi-res_encoder test.
Marco Paniconi [Thu, 10 Jan 2013 02:26:22 +0000 (18:26 -0800)]
Change some settings in multi-res_encoder test.

Change-Id: Ic6a0c3a4b3bb00584d1e9dcdda0b50bdebab87aa

11 years agoMerge superblocks64 experiment.
Ronald S. Bultje [Thu, 10 Jan 2013 01:21:28 +0000 (17:21 -0800)]
Merge superblocks64 experiment.

Change-Id: If6c88752dffdb566f8d4322f135145270716fb8e

11 years agoexperiment a wider loop filter for MB border
Yaowu Xu [Wed, 19 Dec 2012 23:53:03 +0000 (15:53 -0800)]
experiment a wider loop filter for MB border

when larger transforms are used

Change-Id: I25251442b44bf251df4c25a1c1fcf71fb2ad913b

11 years agoNew prediction filter
Adrian Grange [Tue, 8 Jan 2013 22:14:01 +0000 (14:14 -0800)]
New prediction filter

This patch removes the old pred-filter experiment and replaces it
with one that is implemented using the switchable filter framework.

If the pred-filter experiment is enabled, three interopolation
filters are tested during mode selection; the standard 8-tap
interpolation filter, a sharp 8-tap filter and a (new) 8-tap
smoothing filter.

The 6-tap filter code has been preserved for now and if the
enable-6tap experiment is enabled (in addition to the pred-filter
experiment) the original 6-tap filter replaces the new 8-tap smooth
filter in the switchable mode.

The new experiment applies the prediction filter in cases of a
fractional-pel motion vector. Future patches will apply the filter
where the mv is pel-aligned and also to intra predicted blocks.

Change-Id: I08e8cba978f2bbf3019f8413f376b8e2cd85eba4

11 years agomake: support V=1 as short form of verbose=yes
Luca Barbato [Sun, 6 Jan 2013 18:19:43 +0000 (19:19 +0100)]
make: support V=1 as short form of verbose=yes

make V=1 is commonly used for the purpose of trigger verboseness
across projects.

Change-Id: Id9dfb4a7eea8338b366ae8dadb56ab02ceabbb58

11 years agoAdds 64x64 hybrid dct/dwt transform
Deb Mukherjee [Tue, 8 Jan 2013 20:18:16 +0000 (12:18 -0800)]
Adds 64x64 hybrid dct/dwt transform

This is to add to the 64x64 transform experiment as an alternative to
a 64x64 DCT.
Two levels of wavelet decomposition is used on a 64x64 block, followed
by 16x16 DCT on the four lowest subbands. The highest three subbands
are left untransformed after the first level DWT.

Change-Id: I3d48d5800468d655191933894df6b46e15adca56

11 years agoMerge "Merge superblocks (32x32) experiment." into experimental
Ronald S. Bultje [Tue, 8 Jan 2013 21:31:37 +0000 (13:31 -0800)]
Merge "Merge superblocks (32x32) experiment." into experimental

11 years agoMerge "vp9_sub_pixel_variance16x2 SSE2 optimization" into experimental
Yunqing Wang [Tue, 8 Jan 2013 20:59:08 +0000 (12:59 -0800)]
Merge "vp9_sub_pixel_variance16x2 SSE2 optimization" into experimental

11 years agoMerge superblocks (32x32) experiment.
Ronald S. Bultje [Tue, 8 Jan 2013 18:29:22 +0000 (10:29 -0800)]
Merge superblocks (32x32) experiment.

Change-Id: I0df99742029834a85c4933652b0587cf5b6b2587

11 years agovp9_sub_pixel_variance16x2 SSE2 optimization
Yunqing Wang [Tue, 8 Jan 2013 18:44:19 +0000 (10:44 -0800)]
vp9_sub_pixel_variance16x2 SSE2 optimization

About 5% decoder speedup.

Change-Id: Ib6687d337af758a536a0e7e289f400990f1f9794

11 years agoMerge vp9-preview changes into experimental branch
John Koleszar [Tue, 8 Jan 2013 18:11:26 +0000 (10:11 -0800)]
Merge vp9-preview changes into experimental branch

Incorportate vp9-preview changes by merging master branch into experimental.

Conflicts:
test/test.mk
vp9/common/vp9_filter.c
vp9/common/vp9_idctllm.c
vp9/common/vp9_invtrans.h
vp9/common/vp9_mbpitch.c
vp9/common/vp9_rtcd_defs.sh
vp9/common/vp9_systemdependent.h
vp9/common/vp9_type_aliases.h
vp9/common/x86/vp9_asm_stubs.c
vp9/common/x86/vp9_subpixel_mmx.asm
vp9/decoder/vp9_decodframe.c
vp9/decoder/vp9_dequantize.c
vp9/decoder/vp9_dequantize.h
vp9/decoder/vp9_onyxd_int.h
vp9/encoder/vp9_bitstream.c
vp9/encoder/vp9_encodeframe.c
vp9/encoder/vp9_rdopt.c

Change-Id: I17f51c3666d1b59cf1a699f87607cbc5d30a87c5

11 years agoMerge "Remove cpi parameter from update_mvcount()"
Scott LaVarnway [Tue, 8 Jan 2013 18:04:21 +0000 (10:04 -0800)]
Merge "Remove cpi parameter from update_mvcount()"

11 years agoreset segement map on key frame
Yaowu Xu [Fri, 14 Dec 2012 07:53:11 +0000 (23:53 -0800)]
reset segement map on key frame

This is to fix a decoder crash when decoder skips a number of frame to
continue decoding from a later key frame.

Change-Id: I3ba116eba6c3440e0528a21f53745f694302e4ad

11 years agoMerge "minor loop filter refactoring and cleanup" into experimental
Yaowu Xu [Tue, 8 Jan 2013 16:40:03 +0000 (08:40 -0800)]
Merge "minor loop filter refactoring and cleanup" into experimental

11 years agominor loop filter refactoring and cleanup
Yaowu Xu [Wed, 19 Dec 2012 19:34:49 +0000 (11:34 -0800)]
minor loop filter refactoring and cleanup

This commit did a couple of minor cleanup/refactoring to prepare for
futher loop filter experiments. It merged y_only version of loop filter
function into the regular one, which makes sure that same logic is used
for functions for picking level and for actual loop filtering.

Change-Id: Id10c94dccd45f58e5310bacfdf6ee63cbb60b86f

11 years agoRemove a few redundant function arguments in encodeframe.c.
Ronald S. Bultje [Mon, 7 Jan 2013 19:41:49 +0000 (11:41 -0800)]
Remove a few redundant function arguments in encodeframe.c.

Also reindent a block of code that was misindented after addition of
the tx32x32 experiment.

Change-Id: Ic3e4aae3effd8a40136da68c9f382af03632ba08

11 years agoRe-enable support for static_threshold (encode_breakout).
Ronald S. Bultje [Mon, 7 Jan 2013 19:02:14 +0000 (11:02 -0800)]
Re-enable support for static_threshold (encode_breakout).

Change-Id: Ibd7380f478d3127f9db91d0a4fd2fd0dfde961ab

11 years agoDon't use tx32x32 for macroblocks.
Ronald S. Bultje [Fri, 4 Jan 2013 20:05:40 +0000 (12:05 -0800)]
Don't use tx32x32 for macroblocks.

Change-Id: Ib674e0153ca360867ab7a20ba291ac9171a01250

11 years ago64x64 blocksize support.
Ronald S. Bultje [Sun, 6 Jan 2013 02:20:25 +0000 (18:20 -0800)]
64x64 blocksize support.

3.2% gains on std/hd, 1.0% gains on hd.

Change-Id: I481d5df23d8a4fc650a5bcba956554490b2bd200

11 years agoFix mode selection infinite loop bug
Adrian Grange [Fri, 4 Jan 2013 17:00:47 +0000 (09:00 -0800)]
Fix mode selection infinite loop bug

Mode selection for SBs could enter an infinite loop because
the interpolation filter mode index was not being reset
correctly.

Change-Id: I4bbe726f29ef5b6836e94884067c46084713cc11

11 years agoFurther change to mv reference search.
Paul Wilkins [Thu, 3 Jan 2013 15:14:36 +0000 (15:14 +0000)]
Further change to mv reference search.

This experimental change reorders the search so
that all possible references that match the target
reference frame are tested first and these in order
of distance from the current block. These will usually
be the highest scoring candidates.

If we do not find enough good candidates this way
we try non matching cases. These will usually be lower
scoring candidates.

The change in order together with breakouts when
we have found enough candidates should reduce
the computational cost and especially reduce the number
of sort operations.

Quality Results:
Std Hd +0.228%, Hd +0.074%, YT +0.046%, derf +0.137%

This effect is probably due to the fact that more distant
weak candidates are now less likely to get "promoted" over
near candidates even if they are repeated.

Change-Id: Iec37e77d88a48ad0ee1f315b14327a95d63f81f6

11 years agoUse seg/ref/mode offsets in loop_filter_partial.
Marco Paniconi [Thu, 3 Jan 2013 23:47:34 +0000 (15:47 -0800)]
Use seg/ref/mode offsets in loop_filter_partial.

Use the proper seg/mode/ref filter offsets when selecting the
frame loop filter level for fast mode (pick_filter_level_fast).

Change-Id: I2473e2131c800ad19755cb6211ad735fecfe2ac0

11 years agoMerge "make cost_coeffs() and tokenize_b() consistent" into experimental
Yaowu Xu [Thu, 3 Jan 2013 17:57:07 +0000 (09:57 -0800)]
Merge "make cost_coeffs() and tokenize_b() consistent" into experimental

11 years agoMerge "Merge cost_coeffs_2x2() into cost_coeffs()" into experimental
Yaowu Xu [Thu, 3 Jan 2013 17:33:21 +0000 (09:33 -0800)]
Merge "Merge cost_coeffs_2x2() into cost_coeffs()" into experimental